CEM3310 equivalent circuit

oldcrow · Post by **oldcrow** » Tue Apr 23, 2013 7:17 pm

Well, I have a working bread-boarded circuit. It uses JH's inverse-parallel exponential multipliers and traditional analog MUXing (and a 555 for peak threshold detection). Trying to make it compact enough to fit two of them on my Oberheim OBX voice card layout. Since the CEM3310 used tiny CVs with input dividers cutting -5V to -~50mV for time parameters, and the OBX CV storage buffers have inverting active attenuators to make this -50mV external to the CEM3310s, I have to use inverting amplifiers locally--not quite the traditional CV input scheme, but in order to avoid having to hack the OBX CV buffers to accept new voice cards this is the best solution. I guess I blame Doug Curtis for the negative-going time CVs, probably to save space in the interdesign chip array the 3310 was born from. ;) More later.

moogah · Post by **moogah** » Tue Apr 23, 2013 7:30 pm

very cool! The oberheim OB-1 used these for the ADSR's on the later models and I've always _loved_ the envelopes on that synth, the reponse seemed very non-linear, without every actually checking it out I always assumed they were exponential rise and log decay... but your comments about the narrow cv range might actually have more to do with it, I remember the adjustment being oddly sensitive but at the same time very useful musically!

If you get this to work I've actually got a rough pcb layout done for the OB-1... a fantastic monosynth with full patch storage! Even better it actually interfaces with modern microcontroller setups very easily.. hmmmmm

negativspace · Post by **negativspace** » Wed Apr 24, 2013 7:05 am

moogah wrote: If you get this to work I've actually got a rough pcb layout done for the OB-1... a fantastic monosynth with full patch storage! Even better it actually interfaces with modern microcontroller setups very easily.. hmmmmm

Oh man, of all the synths I miss most it's the OB-1. Smoothest, most buttery low-end I ever had in the studio. I've been trying to reacquire one for years... this sounds interesting!

(Sorry, that has nothing to do with the 3310.)

filterstein · Post by **filterstein** » Wed Apr 24, 2013 12:35 pm

Two great ideas in one thread!

My ob1 has discrete envelopes btw.

Dave Kendall · Post by **Dave Kendall** » Thu Apr 25, 2013 3:23 am

moogah wrote:

I've actually got a rough pcb layout done for the OB-1... a fantastic monosynth with full patch storage!

Now THAT is really something.....

Definitely be interested. Any more details Jeff?
Basic MIDI, eg. note in and PGM change to patch select would be awesome...

DIY and patch storage. Yumm. Very tempting....

cheers,
Dave

moogah · Post by **moogah** » Thu Apr 25, 2013 11:18 am

Apologies for the thread-jacking, but I suppose this is on topic enough

Making an new OB-1 is something I could definitely pursue, but I don't think I'll be actively working on it until this fall in any case. My general plan is to create a voice pcb with a layout similar to the 9090 so that all off board connections are on headers organized at the edge of the PCB. A seperate control board can then be created which could either be just pure simple analog with pots or a full on digital setup.

The midibox hardware is already a perfect fit, literally using the same signal levels and multiplexers as the original

the apparent scarcity of the 3080 is a concern though, there are many of them in the circuit and I highly doubt it will really sound correct without them!

Another interesting note is how similar the OB-1 is to the SH-2 in design

.. I've got a layout in the works for that one too... but for real, I'm not trying to clone every old box out there LOL.

glmodular · Post by **glmodular** » Thu Apr 25, 2013 12:39 pm

Scott,

Great news on a 3310 replacement!

Keep me posted as I would dearly love to have these for my own OB-X!

RE: OB-1

I repaired one of these once. Deemed unrepairable by at least two other techs. Found out one VCO had a bunch of transistors replaced with ones from a different manufacturer that were NOT pin compatible!

Very nice synth though! If they had poly'd that configuration, it would have been the ultimate Oberheim!

Tony

Dave Kendall · Post by **Dave Kendall** » Thu Apr 25, 2013 1:03 pm

Hope I'm not adding to the hijacking here

......

the apparent scarcity of the 3080 is a concern though, there are many of them in the circuit and I highly doubt it will really sound correct without them

According to This Link by its designer, the LM13600 is pretty much a straight cut and paste of two 3080s onto a single IC. There are some tiny differences between the 13600 and 13700, IIRC there has been some talk on the [sdiy] list about exactly what they are.

Out of interest, any particular MIDIbox project you were thinking of for control? I'm a bit behind with what uCapps is up to these days....

cheers,
Dave

oldcrow · Post by **oldcrow** » Thu Apr 25, 2013 4:56 pm

Heya Tony,

Jurgen had this figured out 15 years ago, I just decided to add it to my "make an equivalent circuit" list after hearing MC's OBX and deciding I need to make enough of one to get that sound myself. More later.

glmodular wrote:Scott,

Great news on a 3310 replacement! :hail: Keep me posted as I would dearly love to have these for my own OB-X!

oldcrow · Post by **oldcrow** » Sat May 04, 2013 12:40 am

For those interested, here is the CEM3310 equivalent circuit I am using in the OBX voice card. It is largely made from standard EG concepts, the only novel part is Jurgen's inverse-parallel expo multipliers as noted at
http://www.jhaible.com/tonline_stuff/hj_vc_hadsr.html

The 555 and 4052 are standard-issue in DIY EGs, Tom Gamble (EFM) and Thomas Henry are just a couple of names I recalled while breadboarding the circuit. The only extra trick I have to use is biasing the 4052's Vee pin slightly negative as the decay and release CVs are 0 to -250mV as per the CEM3310's time CV input spec, and 405x chips tend to latch up and explode if they see a voltage below a diode drop or so with respect to Vee for analog signals or Vdd for logic signals. Clamp diodes on the analog inputs are also used for this reason.

The 10-Ohm resistors are from the OBX voice card, probably something to do with trimming the maximum times that they never quite got around to. Since the OBX parameter CV storage buffers on their control board use inverting attenuators to take the 0 to +5V down to 0 to -250mV, I have to restore the full scale CV range using IC1, otherwise I would have to modify the OBX control board and for a drop-in voice card not particularly desirable. The trimmers are for figuring out the ideal resistors to actually use for the scaling--also taken from JH's HADSR.

--Crow

Isaiah · Post by **Isaiah** » Sun May 05, 2013 5:58 pm

moogah wrote:Apologies for the thread-jacking, but I suppose this is on topic enough

Making an new OB-1 is something I could definitely pursue, but I don't think I'll be actively working on it until this fall in any case. My general plan is to create a voice pcb with a layout similar to the 9090 so that all off board connections are on headers organized at the edge of the PCB. A seperate control board can then be created which could either be just pure simple analog with pots or a full on digital setup.

The midibox hardware is already a perfect fit, literally using the same signal levels and multiplexers as the original

the apparent scarcity of the 3080 is a concern though, there are many of them in the circuit and I highly doubt it will really sound correct without them!

Another interesting note is how similar the OB-1 is to the SH-2 in design .. I've got a layout in the works for that one too... but for real, I'm not trying to clone every old box out there LOL.

I'll be keeping my eyes open for any progress you make with the OB-1 clone later in the year

oldcrow · Post by **oldcrow** » Tue Sep 10, 2013 4:41 pm

Updated from my crOwBX rev2 tweaks,

alain6870 · Post by **alain6870** » Wed Sep 11, 2013 11:08 am

Cool design when you handle 0 to 5V CVs.

Must the PNP and NPN pairs be macthed?

The Real MC · Post by **The Real MC** » Wed Sep 11, 2013 1:12 pm

I don't think matching is necessary, they feed charging current to C1 which is where the transients are generated. Tempco errors (very small) only impact the A/D/R transients, and the human ear isn't as sensitive to shifts in timing transients as much as shifts in tuning.

oldcrow · Post by **oldcrow** » Wed Sep 11, 2013 3:47 pm

alain6870 wrote:Cool design when you handle 0 to 5V CVs.

Must the PNP and NPN pairs be macthed?

No matching needed. I've found modern 2N3904s and 2N3906s to be close in Vbe when randomly selected in pairs from the parts bins as it is. For an EG it is not critical anyway.

Massimo Ischia · Post by **Massimo Ischia** » Wed Nov 20, 2013 6:44 pm

Scott, your ADSR schematic is very similar to the one I developed in the 2004, see :

http://digilander.libero.it/ismsynth/vc-adsr-1.2.jpg

The core is the well known Jurgen Haible expo converter, but I got the idea of comparing the EG output with the capacitor asymptotic voltage and use the difference as a reference current for the expo.
I am currently trying to put the whole circuit in a small SMD module, not an exact CEM3310 replacement, but something small and useable as a building block for a polyphonic synth.
Bye

oldcrow · Post by **oldcrow** » Wed Nov 20, 2013 7:16 pm

Huh, interesting. If I'd seen this schematic six months ago it would have saved a bit of time.

I basically took JH's VC HADSR and reworked it with the traditional elements as used by Tom Gamble, Thomas Henry and possibly other guys named Tom.

I put it to good use: http://www.cs80.com/crowbx

--Crow
/**/

Massimo Ischia wrote:Scott, your ADSR schematic is very similar to the one I developed in the 2004, see :

http://digilander.libero.it/ismsynth/vc-adsr-1.2.jpg

gdavis · Post by **gdavis** » Wed Nov 20, 2013 8:54 pm

This is pretty cool.

I'm working on a poly synth right now and spent a lot of time looking for a suitable VCADSR. At the moment I'm using JH's VC HADSR with the H part removed but can't help wondering if there would be an advantage to using one of the circuits in this thread?

Massimo Ischia · Post by **Massimo Ischia** » Thu Nov 21, 2013 6:59 am

I prefer to work with IC rather than discrete transistors, so I prefer my and Scott schematics rather than the JH one.
Among other things, our solution is more clean (neat ?) and the operation is easier to understand (IMHO). The difference between my schematics and that of Scott is that my accepts control voltages between 0 and 5V (but it is preferable to place a buffer or use the output of an SH before the control inputs). Scott's one accepts negative voltages as for the CEM3310.

The Real MC · Post by **The Real MC** » Thu Nov 21, 2013 7:07 am

Massimo Ischia wrote:The difference between my schematics and that of Scott is that my accepts control voltages between 0 and 5V (but it is preferable to place a buffer or use the output of an SH before the control inputs). Scott's one accepts negative voltages as for the CEM3310.

Actually if you remove the attenuator/inverter on the A/D/R lines then the circuit responds to 0-5V range. Scott added them so his CrOwBX voicecard would be drop-in compatible with vintage OBX. His schemo notes this:

oldcrow wrote:Updated from my crOwBX rev2 tweaks,

gdavis · Post by **gdavis** » Thu Nov 21, 2013 6:25 pm

It's definitely cleaner and easier to follow. When I saw this my heart kind of sank a bit because I already did a board layout for the JH ADSR. I'm just about to order a whole batch of 16 boards and contemplating whether I should redo it with this implementation or just stick with what I have. But since it's done and works, probably not worth the extra effort at this point.

Littlem4tt · Post by **Littlem4tt** » Sun Apr 12, 2015 7:17 pm

This looks awesome, i look forward to trying it out on the old bread board.
great work all round!

oldcrow · Post by **oldcrow** » Mon Jun 06, 2016 2:05 am

I will have a limited run of my rev 1 bare boards for what I call the 'model 331' envelope generator available soon. This is a 10HP euro format skiff-friendly EG made from this same circuit. I will post the .fpd file for a panel as well.

More soon

--Crow

nvining · Post by **nvining** » Mon Jun 06, 2016 1:33 pm

You are just on fire this summer!

oldcrow · Post by **oldcrow** » Mon Jun 06, 2016 11:37 pm

Here is what rev 1 looks like, FYI

MUFF WIGGLER

MUFF WIGGLER

CEM3310 equivalent circuit

CEM3310 equivalent circuit